INDEX

Explanations

phrases expressing speculative or counterfactual situations

oai_token-act-pair · gpt-3.5-turbo

the phrase "would have" and its variations

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

 muse

-0.63

 rumor

-0.60

 narrator

-0.59

 plaintiff

-0.57

 disinfect

-0.55

 doll

-0.54

 Cold

-0.54

 trope

-0.54

 reminder

-0.53

 marqu

-0.53

POSITIVE LOGITS

 been

1.03

 gotten

0.93

been

0.93

¶

0.89

ĸļ

0.89

Ģ

0.80

 taken

0.79

ãĥ´ãĤ¡

0.78

 Been

0.78

gotten

0.76

Activations Density 0.076%

phrases expressing speculative or counterfactual situations

the phrase "would have" and its variations

No Comments

No Known Activations