INDEX

Explanations

mentions of individuals with the initials "MR"

oai_token-act-pair · gpt-3.5-turbo

references to a specific entity denoted by "MR" or related titles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

éĹĺ

-0.73

stall

-0.66

âĸ¬

-0.64

 Staten

-0.64

tin

-0.63

 Corpus

-0.63

 Hots

-0.62

spring

-0.62

)]

-0.62

 Strait

-0.61

POSITIVE LOGITS

acket

0.96

andom

0.83

angelo

0.81

agnar

0.81

udge

0.78

hammad

0.77

 Lumpur

0.76

CTV

0.76

adish

0.75

idges

0.75

Activations Density 0.005%

mentions of individuals with the initials "MR"

references to a specific entity denoted by "MR" or related titles

No Comments

No Known Activations