INDEX

Explanations

mentions of numerical quantities or statistics

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "One" emphasizing its significance 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 @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 actionGroup

-0.77

cy

-0.72

lov

-0.67

vel

-0.62

Ã¦

-0.61

ooks

-0.60

sed

-0.58

aqu

-0.58

inas

-0.57

ãĥ¼

-0.57

POSITIVE LOGITS

 Hundred

0.98

ĪĴ

0.87

 Thousand

0.80

Voice

0.79

Way

0.77

bestos

0.76

 Piece

0.76

Drive

0.75

auga

0.73

 Million

0.72

Activations Density 0.034%

mentions of numerical quantities or statistics

occurrences of the word "One" emphasizing its significance in various contexts

No Comments

No Known Activations