INDEX

Explanations

phrases related to giving recognition or showing admiration, often in the form of tributes

oai_token-act-pair · gpt-3.5-turbo

references to tributes or dedications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

ī

-0.75

ucket

-0.71

thing

-0.69

handle

-0.69

ECK

-0.63

zee

-0.63

Handle

-0.62

wer

-0.62

wick

-0.61

Techn

-0.61

POSITIVE LOGITS

giving

0.88

 honoring

0.78

oise

0.78

 honors

0.76

 album

0.76

ribute

0.73

ission

0.72

 altar

0.72

warming

0.71

 plaque

0.70

Activations Density 0.025%

phrases related to giving recognition or showing admiration, often in the form of tributes

references to tributes or dedications

No Comments

No Known Activations