INDEX

Explanations

words related to all-encompassing actions or scenarios

oai_token-act-pair · gpt-3.5-turbo

the word "all" and its various contexts in the document

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

SHIP

-0.68

IND

-0.64

gee

-0.63

 Kamp

-0.61

WF

-0.60

lic

-0.60

chie

-0.59

fman

-0.59

bal

-0.58

ritic

-0.58

POSITIVE LOGITS

 kinds

1.50

 sorts

1.48

igators

1.24

igator

1.01

usions

1.00

ocating

0.98

udes

0.91

 manner

0.91

 these

0.87

edged

0.86

Activations Density 0.092%

words related to all-encompassing actions or scenarios

the word "all" and its various contexts in the document

No Comments

No Known Activations