INDEX

Explanations

terms related to exceeding limits or boundaries

oai_token-act-pair · gpt-3.5-turbo

forms of the word "over" in connection with various actions or conditions

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

Brave

-0.76

minster

-0.73

iframe

-0.69

 Seat

-0.65

Pru

-0.64

PO

-0.62

agne

-0.61

 Colo

-0.61

ICLE

-0.61

Vin

-0.60

POSITIVE LOGITS

adows

0.81

itized

0.71

xual

0.68

uled

0.68

ukong

0.65

ealous

0.64

agog

0.64

erences

0.64

 sensit

0.64

whel

0.64

Activations Density 0.073%

terms related to exceeding limits or boundaries

forms of the word "over" in connection with various actions or conditions

No Comments

No Known Activations