INDEX

Explanations

negative conditional statements

oai_token-act-pair · gpt-3.5-turbo

negations and expressions of doubt or hesitation

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

Reviewer

-0.80

 Advice

-0.79

ãĥ³ãĤ¸

-0.75

©¶æ¥µ

-0.74

ingers

-0.74

ãĥ¼ãĥĨãĤ£

-0.74

çīĪ

-0.73

terness

-0.73

ÃįÃį

-0.71

ļé

-0.69

POSITIVE LOGITS

 suffice

0.74

 properly

0.73

 already

0.71

eret

0.71

 otherwise

0.71

 comply

0.71

 succeed

0.69

 exist

0.66

hess

0.66

hin

0.66

Activations Density 0.112%

negative conditional statements

negations and expressions of doubt or hesitation

No Comments

No Known Activations