INDEX

Explanations

mentions of individuals and their actions or characteristics

oai_token-act-pair · gpt-3.5-turbo

elements related to sports and notable achievements

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

 goddamn

-0.69

itized

-0.65

Fuck

-0.62

]]

-0.60

hetic

-0.60

...)

-0.59

Mods

-0.58

.[

-0.58

Quantity

-0.58

âĢ¦)

-0.58

POSITIVE LOGITS

 criticised

0.89

 controvers

0.78

 angered

0.72

 ousted

0.72

 spokesman

0.72

 sacked

0.70

 embroiled

0.68

 outgoing

0.67

 hailed

0.66

 controversial

0.66

Activations Density 1.475%

mentions of individuals and their actions or characteristics

elements related to sports and notable achievements

No Comments

No Known Activations