INDEX

Explanations

words related to specific names or identities

oai_token-act-pair · gpt-3.5-turbo

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

oses

-0.76

istics

-0.64

astically

-0.59

ournal

-0.58

ANK

-0.58

ials

-0.57

iary

-0.56

astic

-0.56

 ______

-0.55

iates

-0.55

POSITIVE LOGITS

lla

1.25

llan

1.17

lli

1.07

llers

1.05

lling

1.05

lda

1.04

ll

1.01

ller

1.00

tta

0.97

hart

0.96

Activations Density 0.125%

words related to specific names or identities

No Comments

No Known Activations