INDEX

Explanations

names or titles, especially those with the word "Baron"

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names and titles associated with individuals

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

mable

-0.79

DN

-0.72

 extrem

-0.67

FORMATION

-0.65

agents

-0.63

mitting

-0.63

************

-0.63

release

-0.63

 skelet

-0.62

EMS

-0.62

POSITIVE LOGITS

 Baron

1.04

ess

1.00

 Mord

0.94

esses

0.93

fman

0.85

stein

0.82

ham

0.82

 Rothschild

0.78

esis

0.78

alist

0.77

Activations Density 0.010%

names or titles, especially those with the word "Baron"

proper nouns, particularly names and titles associated with individuals

No Comments

No Known Activations