INDEX

Explanations

proper nouns, possibly related to business or politics

oai_token-act-pair · gpt-3.5-turbo

instances of a particular symbol or character in the context of various statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

izen

-0.87

 Slug

-0.65

agall

-0.63

 barrage

-0.61

chnology

-0.60

imeter

-0.60

ggy

-0.59

enium

-0.58

 Orchestra

-0.57

 Gael

-0.57

POSITIVE LOGITS

which

1.29

including

1.18

particularly

1.17

then

1.12

where

1.11

especially

1.11

except

1.10

yet

1.08

thus

1.05

according

1.01

Activations Density 0.179%

proper nouns, possibly related to business or politics

instances of a particular symbol or character in the context of various statements

No Comments

No Known Activations