INDEX

Explanations

words related to specific locations or entities such as "Atelier" and "Hamas" sitting at the top of activations

oai_token-act-pair · gpt-3.5-turbo

words related to artistic and creative activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 Weasley

-0.60

 Cinderella

-0.60

 winters

-0.59

 Cosmos

-0.59

 Ghostbusters

-0.58

verb

-0.57

 Chao

-0.56

ruary

-0.56

glomer

-0.56

 towels

-0.56

POSITIVE LOGITS

mast

1.04

antage

0.78

 mast

0.76

elia

0.74

llah

0.73

obl

0.72

rophe

0.70

rius

0.69

oma

0.69

igible

0.69

Activations Density 0.055%

words related to specific locations or entities such as "Atelier" and "Hamas" sitting at the top of activations

words related to artistic and creative activities

No Comments

No Known Activations