INDEX

Explanations

names and specific identifying information

oai_token-act-pair · gpt-3.5-turbo

phrases related to names and their significance or identification

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

vable

-0.84

ractical

-0.80

 resil

-0.79

 stalls

-0.78

issance

-0.78

grade

-0.76

issions

-0.75

alysis

-0.74

 teaches

-0.72

edience

-0.70

POSITIVE LOGITS

 initials

0.88

 nationality

0.81

è£

0.74

 pronouns

0.74

 trademarks

0.73

 Uzbek

0.72

 amen

0.72

Tup

0.71

 Arabic

0.69

 Whats

0.69

Activations Density 0.398%

names and specific identifying information

phrases related to names and their significance or identification

No Comments

No Known Activations