INDEX

Explanations

specific names related to health or medical topics

oai_token-act-pair · gpt-3.5-turbo

variations of the word "safa" or related themes

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

 Halls

-0.74

 sidx

-0.72

 inqu

-0.71

 Olympia

-0.70

 Hulk

-0.66

displayText

-0.65

 annexed

-0.60

Pont

-0.60

 Izan

-0.59

 Gemini

-0.59

POSITIVE LOGITS

rican

1.16

ayette

1.08

rica

1.06

avorite

1.01

af

1.01

raid

1.01

onso

1.00

ranch

0.96

ornia

0.96

aina

0.93

Activations Density 0.006%

specific names related to health or medical topics

variations of the word "safa" or related themes

No Comments

No Known Activations