INDEX

Explanations

references to armed individuals or situations involving armed individuals

oai_token-act-pair · gpt-3.5-turbo

mentions of armed forces or individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rx

-1.00

 Remastered

-0.82

rb

-0.77

arget

-0.75

HUD

-0.74

atom

-0.71

article

-0.70

âĸ¬

-0.69

 Decay

-0.68

 coli

-0.67

POSITIVE LOGITS

 robbery

1.08

 guards

0.98

 armed

0.94

 insurrection

0.93

 robber

0.92

 robbers

0.90

 robberies

0.87

 uprising

0.80

ament

0.79

guards

0.78

Activations Density 0.014%

references to armed individuals or situations involving armed individuals

mentions of armed forces or individuals

No Comments

No Known Activations