INDEX

Explanations

phrases indicating actions or states of being

oai_token-act-pair · gpt-3.5-turbo

auxiliary verbs indicating actions or states

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

QB

-0.79

WAR

-0.65

amel

-0.64

 incorpor

-0.62

imoto

-0.62

direction

-0.61

ops

-0.60

otide

-0.59

ibaba

-0.58

imer

-0.58

POSITIVE LOGITS

Geh

0.72

 Sanct

0.70

 Nost

0.68

initions

0.66

 Kern

0.64

vation

0.64

gh

0.60

doi

0.60

ndra

0.59

 ours

0.59

Activations Density 0.051%

phrases indicating actions or states of being

auxiliary verbs indicating actions or states

No Comments

No Known Activations