INDEX

Explanations

words related to leadership and authority, especially with the word "cap" as part of the term

oai_token-act-pair · gpt-3.5-turbo

references to "capital" and its variations in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

hower

-0.80

ĪĴ

-0.67

 Nare

-0.63

WAY

-0.63

 anten

-0.62

 Fein

-0.62

 silence

-0.61

VD

-0.61

 Akin

-0.60

 directional

-0.57

POSITIVE LOGITS

itol

1.55

illary

1.38

rices

1.32

uchin

1.30

rice

1.30

itals

1.21

acity

1.13

abilities

1.11

ric

1.08

itated

1.06

Activations Density 0.027%

words related to leadership and authority, especially with the word "cap" as part of the term

references to "capital" and its variations in various contexts

No Comments

No Known Activations