INDEX

Explanations

the word "seemingly" followed by some description of an event or object

oai_token-act-pair · gpt-3.5-turbo

the word "seemingly" and its variations, indicating uncertainty or appearances

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

ioch

-0.79

ests

-0.73

 trainers

-0.72

iths

-0.66

rike

-0.65

otiation

-0.65

xual

-0.65

iov

-0.64

seller

-0.64

alez

-0.64

POSITIVE LOGITS

 innocuous

0.97

icably

0.90

Buyable

0.82

 unstoppable

0.77

 contradictory

0.77

 unrelated

0.76

 insignificant

0.71

 Tradable

0.71

 limitless

0.71

 contradictions

0.69

Activations Density 0.014%

the word "seemingly" followed by some description of an event or object

the word "seemingly" and its variations, indicating uncertainty or appearances

No Comments

No Known Activations