INDEX

Explanations

dates in the format of day, month, and year

oai_token-act-pair · gpt-3.5-turbo

date references and chronological information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 unpredictable

-0.75

 steroids

-0.73

 imagination

-0.73

 conce

-0.71

gettable

-0.70

avorite

-0.68

 entrepreneurship

-0.67

 unpredict

-0.67

imble

-0.67

 predec

-0.66

POSITIVE LOGITS

 RELEASE

1.13

 Updated

0.99

 ISBN

0.96

GMT

0.92

 Update

0.85

 Apply

0.85

TAG

0.84

 Referred

0.84

 Dear

0.84

doi

0.84

Activations Density 0.063%

dates in the format of day, month, and year

date references and chronological information

No Comments

No Known Activations