INDEX

Explanations

references to bus-related information

oai_token-act-pair · gpt-3.5-turbo

references to "bus" or related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 payoff

-0.73

 marrow

-0.70

INAL

-0.64

arning

-0.63

iamond

-0.62

 consequence

-0.61

ulhu

-0.60

venge

-0.59

bered

-0.59

 realization

-0.58

POSITIVE LOGITS

bus

1.27

INESS

1.04

hel

0.94

peed

0.85

Rus

0.85

Bus

0.85

 Dumbledore

0.80

quin

0.79

cules

0.78

olean

0.78

Activations Density 0.010%

references to bus-related information

references to "bus" or related terms

No Comments

No Known Activations