INDEX

Explanations

adjectives describing something positively or favorably

oai_token-act-pair · gpt-3.5-turbo

the recurring use of the word "pretty" to indicate a positive sentiment or evaluation

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

venant

-0.85

pent

-0.77

arta

-0.72

upon

-0.71

FTA

-0.69

ossession

-0.69

APD

-0.69

allas

-0.68

iments

-0.68

herry

-0.67

POSITIVE LOGITS

 darn

1.30

 nifty

0.98

 damn

0.91

 much

0.88

 harmless

0.87

 neat

0.83

 tasty

0.83

 damned

0.83

 handy

0.82

 cool

0.81

Activations Density 0.018%

adjectives describing something positively or favorably

the recurring use of the word "pretty" to indicate a positive sentiment or evaluation

No Comments

No Known Activations