INDEX

Explanations

corresponding or related items or information

oai_token-act-pair · gpt-3.5-turbo

instances of the word "correspond" and its variations, indicating comparisons or relationships

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

cher

-0.72

uv

-0.71

bane

-0.69

uff

-0.69

sk

-0.68

zy

-0.68

Bay

-0.67

bay

-0.66

spe

-0.66

ubb

-0.66

POSITIVE LOGITS

ivil

0.77

ãĤ±

0.73

ãĥĺ

0.72

 newsp

0.71

 guiActiveUn

0.70

ities

0.69

ively

0.69

 Pengu

0.69

 Occupations

0.69

encies

0.69

Activations Density 0.007%

corresponding or related items or information

instances of the word "correspond" and its variations, indicating comparisons or relationships

No Comments

No Known Activations