INDEX

Explanations

terms related to geographical locations, specifically seashores

oai_token-act-pair · gpt-3.5-turbo

variations of the word "sea" and associated terms

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Pigs

-0.78

Pwr

-0.65

 Bills

-0.64

 clicked

-0.60

 Bennett

-0.58

 Andersen

-0.58

 innocence

-0.58

ãĤ¤ãĥĪ

-0.58

 merry

-0.56

 buffalo

-0.56

POSITIVE LOGITS

fried

0.85

arth

0.84

ĸļ

0.81

orage

0.81

ember

0.79

ecast

0.78

achers

0.75

ached

0.74

eper

0.73

eur

0.72

Activations Density 0.082%

terms related to geographical locations, specifically seashores

variations of the word "sea" and associated terms

No Comments

No Known Activations