INDEX

Explanations

years or dates in a specific format

oai_token-act-pair · gpt-3.5-turbo

locations and names of museums or theaters

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

omething

-0.76

ccording

-0.72

vernment

-0.67

staking

-0.62

ensibly

-0.62

 reconc

-0.59

 scram

-0.59

 sort

-0.58

 scrambling

-0.58

 compromises

-0.56

POSITIVE LOGITS

 ********************************

0.71

<|endoftext|>

0.71

aceae

0.67

pmwiki

0.66

0.65

 UCHIJ

0.65

NX

0.64

 Remix

0.64

 =====

0.63

Activations Density 0.192%

years or dates in a specific format

locations and names of museums or theaters

No Comments

No Known Activations