INDEX

Explanations

statements or opinions beginning with "that"

oai_token-act-pair · gpt-3.5-turbo

the word "that" used 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

 pione

-0.81

tein

-0.73

entary

-0.69

ActionCode

-0.67

Ĝ

-0.66

ptoms

-0.66

ă

-0.66

ü

-0.66

ė

-0.66

 RandomRedditor

-0.66

POSITIVE LOGITS

"[

1.23

he

1.23

 although

1.08

if

1.05

 "âĢ¦

1.04

she

1.04

 unless

0.98

 "...

0.97

 they

0.93

"(

0.91

Activations Density 0.234%

statements or opinions beginning with "that"

the word "that" used in various contexts

No Comments

No Known Activations