INDEX

Explanations

questions ending with a question mark

oai_token-act-pair · gpt-3.5-turbo

rhetorical questions

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 celebr

-0.68

 hust

-0.66

 proud

-0.62

 inactive

-0.62

 cross

-0.60

andi

-0.59

 migration

-0.59

 contagious

-0.58

 happy

-0.57

 upstream

-0.56

POSITIVE LOGITS

Answer

1.51

Well

1.05

ÂłÂłÂłÂł

0.96

Yes

0.96

Probably

0.95

Solution

0.94

YES

0.91

 Answer

0.89

ÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂł

0.89

Correct

0.87

Activations Density 0.141%

questions ending with a question mark

rhetorical questions

No Comments

No Known Activations