INDEX

Explanations

phrases related to political figures and events

oai_token-act-pair · gpt-3.5-turbo

instances of possessive pronouns or indicators of ownership

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

OLOGY

-0.69

Capture

-0.64

EEE

-0.61

ship

-0.59

rolling

-0.59

LESS

-0.56

ATTLE

-0.55

 graft

-0.55

 citation

-0.55

ASED

-0.55

POSITIVE LOGITS

kaya

1.40

omething

1.33

ilver

1.29

poon

1.29

cript

1.28

aurus

1.23

ki

1.23

ullivan

1.22

hip

1.21

hire

1.21

Activations Density 0.209%

phrases related to political figures and events

instances of possessive pronouns or indicators of ownership

No Comments

No Known Activations