INDEX

Explanations

verbs or phrases indicating a sequence of actions or steps

oai_token-act-pair · gpt-3.5-turbo

phrases relating to actions taken by the subject

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Cosponsors

-0.85

 Nanto

-0.73

 Volunteers

-0.72

 Presbyter

-0.69

Pis

-0.68

pse

-0.66

 contrasts

-0.65

 Christy

-0.65

ccording

-0.65

 Preservation

-0.64

POSITIVE LOGITS

uberty

0.96

rosis

0.84

rehend

0.83

irtual

0.81

haul

0.81

ihad

0.80

hran

0.80

acqu

0.79

pection

0.77

verified

0.77

Activations Density 0.349%

verbs or phrases indicating a sequence of actions or steps

phrases relating to actions taken by the subject

No Comments

No Known Activations