INDEX

Explanations

instances where someone is proven wrong

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "wrong."

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

Ri

-0.70

hens

-0.69

 Flavoring

-0.67

incinn

-0.65

coni

-0.62

Fn

-0.61

hedral

-0.59

ats

-0.58

electric

-0.57

kamp

-0.57

POSITIVE LOGITS

fully

0.96

eous

0.95

 guiActiveUn

0.81

headed

0.81

ed

0.74

sight

0.73

 unfocusedRange

0.72

spin

0.70

doing

0.67

dest

0.66

Activations Density 0.018%

instances where someone is proven wrong

occurrences of the word "wrong."

No Comments

No Known Activations