INDEX

Explanations

words related to violence and violations

oai_token-act-pair · gpt-3.5-turbo

terms related to violations and legal issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

dress

-0.77

mares

-0.76

ointment

-0.74

 guiActiveUnfocused

-0.69

erald

-0.68

washing

-0.62

Ħ¢

-0.61

mare

-0.61

nect

-0.61

 wealth

-0.59

POSITIVE LOGITS

viol

1.05

 violin

0.91

atis

0.90

amental

0.88

 Viol

0.88

atos

0.87

encies

0.85

atio

0.83

ace

0.82

isions

0.82

Activations Density 0.017%

words related to violence and violations

terms related to violations and legal issues

No Comments

No Known Activations