INDEX

Explanations

references to a particular acronym "WR" followed by a numerical value

oai_token-act-pair · gpt-3.5-turbo

references to the 'WR' designation, likely indicating a specific type of character or item identifier in a context

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

anamo

-0.80

 Maid

-0.65

uations

-0.63

omorphic

-0.62

onyms

-0.59

 Crosby

-0.58

 Belg

-0.58

 Debor

-0.57

cles

-0.57

ovic

-0.56

POSITIVE LOGITS

ONG

1.18

ACK

1.08

EST

0.96

MJ

0.93

IGHT

0.90

WR

0.88

IVES

0.87

ITT

0.86

ATH

0.86

IGHTS

0.84

Activations Density 0.009%

references to a particular acronym "WR" followed by a numerical value

references to the 'WR' designation, likely indicating a specific type of character or item identifier in a context

No Comments

No Known Activations