INDEX

Explanations

open-ended questions typically starting with "so what" or "what"

oai_token-act-pair · gpt-3.5-turbo

questions or inquiries that prompt further discussion or thought

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

 pione

-0.74

ñ

-0.68

ested

-0.66

anus

-0.66

interstitial

-0.65

ø

-0.65

ā

-0.65

č

-0.65

 RandomRedditor

-0.65

Ğ

-0.64

POSITIVE LOGITS

 exactly

1.15

 happens

1.06

 does

1.02

 SHOULD

0.90

 constitutes

0.90

do

0.88

does

0.88

 recourse

0.88

 qualifies

0.86

 DOES

0.86

Activations Density 0.079%

open-ended questions typically starting with "so what" or "what"

questions or inquiries that prompt further discussion or thought

No Comments

No Known Activations