INDEX

Explanations

mentions of the word "double" and giving it a high activation score

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "double" in various contexts

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

utical

-0.88

Interstitial

-0.88

ULE

-0.86

CHAT

-0.83

utics

-0.80

akeru

-0.80

ä½ľ

-0.79

ihara

-0.78

atche

-0.77

ayn

-0.76

POSITIVE LOGITS

header

0.92

 sided

0.89

 digits

0.88

 digit

0.88

 dipping

0.81

headers

0.81

leaf

0.78

 negatives

0.71

ceptions

0.71

cell

0.69

Activations Density 0.012%

mentions of the word "double" and giving it a high activation score

references to the concept of "double" in various contexts

No Comments

No Known Activations