INDEX

Explanations

words related to misunderstandings or negative actions

oai_token-act-pair · gpt-3.5-turbo

variations of the root word "mis" indicating negative actions or perceptions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 Fres

-0.69

flix

-0.66

itaire

-0.64

 premie

-0.64

 Collider

-0.64

ulhu

-0.61

 disse

-0.60

 cooled

-0.59

retty

-0.57

hower

-0.57

POSITIVE LOGITS

vous

0.97

dule

0.81

ippi

0.78

imates

0.76

gotten

0.76

akable

0.71

Ö

0.71

imate

0.71

ukong

0.70

ãĥ«

0.70

Activations Density 0.039%

words related to misunderstandings or negative actions

variations of the root word "mis" indicating negative actions or perceptions

No Comments

No Known Activations