INDEX

Explanations

references to autographs or items being autographed

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "authograph" and its variations, indicating a focus on signed items or collectibles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

IRO

-0.82

UTION

-0.80

 Soda

-0.80

lihood

-0.79

imaru

-0.76

BE

-0.75

ENCE

-0.74

ptives

-0.74

 Forbidden

-0.72

×Ķ

-0.72

POSITIVE LOGITS

ographs

1.29

ograph

1.21

ographed

1.04

istically

1.03

iques

1.02

ocom

1.00

istic

0.97

umn

0.96

ogyn

0.95

opsy

0.94

Activations Density 0.008%

references to autographs or items being autographed

occurrences of the term "authograph" and its variations, indicating a focus on signed items or collectibles

No Comments

No Known Activations