INDEX

Explanations

words related to strong language or intensity

oai_token-act-pair · gpt-3.5-turbo

instances of the word "rhetoric."

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

ITNESS

-0.68

TAIN

-0.63

IER

-0.62

essa

-0.60

atch

-0.60

ECK

-0.60

atar

-0.60

emale

-0.59

cot

-0.58

cop

-0.58

POSITIVE LOGITS

 rhetoric

0.95

 flare

0.79

mith

0.76

 emanating

0.76

 flared

0.73

 surrounding

0.73

tir

0.71

 denouncing

0.71

 spew

0.71

 flares

0.70

Activations Density 0.026%

words related to strong language or intensity

instances of the word "rhetoric."

No Comments

No Known Activations