INDEX

Explanations

names and titles of political figures

oai_token-act-pair · gpt-3.5-turbo

names of individuals and references to their affiliations or roles

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ĸļ

-0.86

 Grateful

-0.80

sylv

-0.79

 Starcraft

-0.78

 humans

-0.75

 mammalian

-0.73

fec

-0.71

 Californ

-0.70

 cryst

-0.69

 veterin

-0.69

POSITIVE LOGITS

 Ahmad

1.18

 Hussein

1.10

uala

1.10

 Shah

1.10

 Hasan

1.10

 Sharif

1.07

 Ibrahim

1.06

 Mohamed

1.05

 Hassan

1.05

ji

1.05

Activations Density 0.269%

names and titles of political figures

names of individuals and references to their affiliations or roles

No Comments

No Known Activations