INDEX

Explanations

instances where something happens or is true in spite of something else

oai_token-act-pair · gpt-3.5-turbo

instances of the word "despite."

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

ahime

-0.81

isa

-0.74

allo

-0.71

ISE

-0.71

isites

-0.70

ault

-0.69

eg

-0.66

oyd

-0.65

apeake

-0.64

oya

-0.63

POSITIVE LOGITS

ĸļ

0.98

 acknowledging

0.85

 admitting

0.80

 insisting

0.80

 having

0.78

 denying

0.74

 ignoring

0.72

 knowing

0.72

 pledging

0.69

 mentioning

0.69

Activations Density 0.028%

instances where something happens or is true in spite of something else

instances of the word "despite."

No Comments

No Known Activations