INDEX

Explanations

references to different entrances, including exits

oai_token-act-pair · gpt-3.5-turbo

references to entrances and exits

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

nian

-0.70

rior

-0.66

uku

-0.65

amera

-0.65

nutrition

-0.64

ilot

-0.62

auld

-0.62

yan

-0.62

akia

-0.62

conn

-0.61

POSITIVE LOGITS

 thereto

0.99

 entrance

0.96

 entrances

0.94

prise

0.93

door

0.82

 doors

0.82

 pupil

0.81

 gates

0.79

 portals

0.78

prises

0.78

Activations Density 0.018%

references to different entrances, including exits

references to entrances and exits

No Comments

No Known Activations