INDEX

Explanations

phrases that denote a concept or something specific

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the concept of "past events or things."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Respons

-0.61

tails

-0.60

burse

-0.60

opus

-0.59

oons

-0.59

CPU

-0.59

oglu

-0.58

abad

-0.58

\-

-0.58

ped

-0.58

POSITIVE LOGITS

 significance

0.90

 importance

0.84

 beauty

0.82

 legend

0.80

 substance

0.78

 nightmares

0.75

 note

0.75

ificantly

0.74

 utmost

0.74

 folklore

0.72

Activations Density 0.034%

phrases that denote a concept or something specific

phrases indicating the concept of "past events or things."

No Comments

No Known Activations