INDEX

Explanations

phrases indicating a new understanding or insight on a specific topic or issue

oai_token-act-pair · gpt-3.5-turbo

phrases related to shedding light on various topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

okia

-0.77

ocide

-0.69

halla

-0.68

ridor

-0.66

 Rampage

-0.66

umbers

-0.65

keye

-0.64

 Niagara

-0.63

 apologize

-0.62

keyes

-0.62

POSITIVE LOGITS

enment

0.90

weights

0.89

lights

0.79

ener

0.79

nings

0.78

 bulb

0.78

hearted

0.75

 shone

0.75

 clues

0.75

forward

0.75

Activations Density 0.015%

phrases indicating a new understanding or insight on a specific topic or issue

phrases related to shedding light on various topics

No Comments

No Known Activations