INDEX

Explanations

violence-related terminology and mentions of military or armed forces

oai_token-act-pair · gpt-3.5-turbo

references to military and state forces involved in conflicts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Hop

-0.84

\\\\\\\\

-0.79

Dub

-0.77

STER

-0.73

âĶģ

-0.73

 Bout

-0.72

CAST

-0.72

Bird

-0.69

Brew

-0.68

MAL

-0.68

POSITIVE LOGITS

 stationed

0.99

 forces

0.92

 force

0.90

 fatig

0.90

 deployed

0.90

 mobilized

0.86

 troops

0.85

 garrison

0.84

forces

0.83

 recruited

0.82

Activations Density 0.018%

violence-related terminology and mentions of military or armed forces

references to military and state forces involved in conflicts

No Comments

No Known Activations