INDEX

Explanations

names ending in 'ler'

oai_token-act-pair · gpt-3.5-turbo

words related to legal or criminal actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

raints

-0.72

ADRA

-0.63

 Occupations

-0.63

Latin

-0.59

ourning

-0.59

ANA

-0.59

popular

-0.58

 Bahrain

-0.58

UAL

-0.57

Its

-0.56

POSITIVE LOGITS

gren

0.96

oad

0.95

iffe

0.89

sonian

0.88

icum

0.86

istics

0.85

ogue

0.84

llan

0.83

mann

0.78

onnaissance

0.78

Activations Density 0.018%

names ending in 'ler'

words related to legal or criminal actions

No Comments

No Known Activations