INDEX

Explanations

objects or things mentioned in a text

oai_token-act-pair · gpt-3.5-turbo

mentions of "object" 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

 millenn

-0.90

ould

-0.66

cling

-0.65

 Sussex

-0.64

ooks

-0.64

 sanity

-0.64

artney

-0.64

 Marin

-0.63

 Hampton

-0.62

rosso

-0.62

POSITIVE LOGITS

handler

0.93

object

0.88

 object

0.88

 objects

0.82

objects

0.81

ifier

0.79

ivity

0.79

imus

0.78

 Object

0.76

 Clicker

0.74

Activations Density 0.013%

objects or things mentioned in a text

mentions of "object" in various contexts

No Comments

No Known Activations