INDEX

Explanations

words and phrases that seem to be a mix of non-English characters and potentially mistyped English words or names

oai_token-act-pair · gpt-3.5-turbo

characters from non-Latin alphabets

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

yip

-0.76

-0.75

 reconc

-0.74

displayText

-0.66

ritical

-0.66

 glimpse

-0.65

irtual

-0.64

 answ

-0.64

arton

-0.64

 crossings

-0.63

POSITIVE LOGITS

Ø©

0.85

ÙĨ

0.84

 ÙĪ

0.83

;;;;;;;;;;;;

0.82

Ùĩ

0.76

 istg

0.76

ÏĤ

0.75

 srfAttach

0.74

 è£ıç

0.72

 Ø§ÙĦ

0.70

Activations Density 0.158%

words and phrases that seem to be a mix of non-English characters and potentially mistyped English words or names

characters from non-Latin alphabets

No Comments

No Known Activations