INDEX

Explanations

names, in particular the name "Darrell."

oai_token-act-pair · gpt-3.5-turbo

mentions of individuals named Darrell or Farrell in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

Oxy

-0.76

ablishment

-0.75

awaru

-0.73

aepernick

-0.72

dict

-0.69

akening

-0.66

real

-0.66

onom

-0.65

 Unch

-0.65

arcer

-0.65

POSITIVE LOGITS

arrell

0.99

issance

0.88

||||

0.85

 Issa

0.84

rette

0.80

ragon

0.80

served

0.72

rical

0.69

inav

0.69

rers

0.68

Activations Density 0.025%

names, in particular the name "Darrell."

mentions of individuals named Darrell or Farrell in the text

No Comments

No Known Activations