INDEX

Explanations

numbers and monetary values, especially with decimals

oai_token-act-pair · gpt-3.5-turbo

numerical data and figures related to quantities or statistics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 shopping

-0.65

 succ

-0.63

 direction

-0.60

 fren

-0.60

 puff

-0.59

 subp

-0.58

 tremend

-0.58

hog

-0.57

 sworn

-0.57

 psyche

-0.57

POSITIVE LOGITS

0.87

0.85

0.84

0.83

0.82

0.81

Activations Density 0.158%

numbers and monetary values, especially with decimals

numerical data and figures related to quantities or statistics

No Comments

No Known Activations