INDEX

Explanations

references to explosive devices or weapons

oai_token-act-pair · gpt-3.5-turbo

references to pagans and grenades

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

sburgh

-0.92

balls

-0.83

yll

-0.83

icle

-0.82

esley

-0.82

angelo

-0.79

cream

-0.76

ening

-0.76

yl

-0.74

POSITIVE LOGITS

vernment

0.93

adier

0.77

atos

0.72

vier

0.70

merce

0.66

issors

0.65

ACTED

0.63

aceae

0.63

âĸ¬âĸ¬

0.62

ptions

0.59

Activations Density 0.090%

references to explosive devices or weapons

references to pagans and grenades

No Comments

No Known Activations