INDEX

Explanations

contractions where the contraction includes "n't" indicating negation

oai_token-act-pair · gpt-3.5-turbo

instances of the contraction "doesn't"

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

 Citiz

-0.78

itiz

-0.75

çĶŁ

-0.71

 Mayhem

-0.69

 Britann

-0.66

 Presence

-0.65

oided

-0.65

jected

-0.64

Minor

-0.63

Reviewer

-0.63

POSITIVE LOGITS

 necessarily

1.33

 bother

1.12

 exactly

1.08

icably

0.97

 seem

0.95

 quite

0.93

 deserve

0.92

 hesitate

0.92

 really

0.89

epad

0.87

Activations Density 0.094%

contractions where the contraction includes "n't" indicating negation

instances of the contraction "doesn't"

No Comments

No Known Activations