INDEX

Explanations

words related to misunderstandings and negative actions

oai_token-act-pair · gpt-3.5-turbo

forms of the word "mis" and its derivatives, indicating themes of mistakes or misunderstandings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Fres

-0.75

itaire

-0.68

flix

-0.68

 Ferr

-0.64

ulhu

-0.62

 Cube

-0.62

ilon

-0.59

retty

-0.59

 Chest

-0.58

raltar

-0.58

POSITIVE LOGITS

vous

0.98

gotten

0.85

rued

0.82

akable

0.82

ceptions

0.81

ceived

0.80

ortunate

0.80

imates

0.79

outed

0.77

dule

0.77

Activations Density 0.053%

words related to misunderstandings and negative actions

forms of the word "mis" and its derivatives, indicating themes of mistakes or misunderstandings

No Comments

No Known Activations