INDEX

Explanations

references to violence and related topics, such as discussions, condemnation, and potential solutions

oai_token-act-pair · gpt-3.5-turbo

mentions of violence and its implications in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

sonian

-0.89

ocular

-0.78

é£

-0.78

é¾įå

-0.77

ITNESS

-0.77

dit

-0.71

glas

-0.70

gres

-0.68

acle

-0.67

amina

-0.67

POSITIVE LOGITS

 perpetrated

1.30

 against

1.19

 Against

1.07

 inflicted

1.04

 prevention

0.99

 directed

0.98

against

0.98

 committed

0.95

rained

0.93

 toward

0.93

Activations Density 0.060%

references to violence and related topics, such as discussions, condemnation, and potential solutions

mentions of violence and its implications in various contexts

No Comments

No Known Activations