INDEX

Explanations

terms related to a scientific field or medical condition, specifically focusing on the concept of detecting errors or abnormalities

oai_token-act-pair · gpt-3.5-turbo

terms related to detection and related medical or scientific procedures

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 rage

-0.69

 gamb

-0.67

 Bloom

-0.64

 Ashley

-0.63

 repeated

-0.62

 brewery

-0.61

 airing

-0.61

 rush

-0.60

 GREEN

-0.60

 Johnny

-0.60

POSITIVE LOGITS

ect

4.81

ection

2.34

ected

2.29

ector

2.06

ective

1.70

ECT

1.66

ectomy

1.44

ECTION

1.26

duct

1.10

ract

1.09

Activations Density 0.006%

terms related to a scientific field or medical condition, specifically focusing on the concept of detecting errors or abnormalities

terms related to detection and related medical or scientific procedures

No Comments

No Known Activations