INDEX

Explanations

sources or attributions by the use of the word "courtesy."

oai_token-act-pair · gpt-3.5-turbo

instances of the word "courtesy" related to sources or attribution

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

oshenko

-0.88

akia

-0.82

ako

-0.80

lear

-0.79

olver

-0.75

dimension

-0.75

iated

-0.75

ommod

-0.74

ago

-0.74

vals

-0.72

POSITIVE LOGITS

tesy

0.96

 courtesy

0.83

 Thomson

0.78

giving

0.70

 Credit

0.69

 Boll

0.69

 Olivier

0.68

 Purdue

0.67

 Cullen

0.67

 Brewer

0.66

Activations Density 0.011%

sources or attributions by the use of the word "courtesy."

instances of the word "courtesy" related to sources or attribution

No Comments

No Known Activations