INDEX

Explanations

words related to heroes or heroic acts

oai_token-act-pair · gpt-3.5-turbo

references to heroes or heroic figures

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

aeda

-0.93

ntil

-0.78

ateur

-0.75

imentary

-0.73

creen

-0.70

ighton

-0.69

acular

-0.68

uce

-0.66

gow

-0.66

cerning

-0.66

POSITIVE LOGITS

ically

1.02

ku

0.90

 hero

0.90

ines

0.88

 heroine

0.88

hero

0.85

 heroes

0.84

Hero

0.80

ics

0.80

å£«

0.76

Activations Density 0.027%

words related to heroes or heroic acts

references to heroes or heroic figures

No Comments

No Known Activations