INDEX

Explanations

mentions of violence in various contexts

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "violence" and related phrases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

-0.84

ocular

-0.82

arton

-0.80

dit

-0.76

ramer

-0.71

osition

-0.71

oplan

-0.71

é¾įå

-0.68

itions

-0.68

odes

-0.66

POSITIVE LOGITS

 perpetrated

1.12

 inflicted

0.97

fighting

0.87

 against

0.85

 violence

0.84

quit

0.81

 directed

0.78

Viol

0.77

 prevention

0.77

 erupted

0.77

Activations Density 0.052%

mentions of violence in various contexts

occurrences of the word "violence" and related phrases

No Comments

No Known Activations