INDEX

Explanations

words related to planned or intentional actions

oai_token-act-pair · gpt-3.5-turbo

past tense verbs, particularly those denoting actions or events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Patron

-0.64

 presiding

-0.62

 Prel

-0.62

 payable

-0.60

 Compos

-0.59

 grading

-0.58

Score

-0.58

 boil

-0.56

 Prize

-0.56

 Divide

-0.56

POSITIVE LOGITS

ted

4.64

ting

3.03

ts

2.00

ned

1.74

tted

1.72

TED

1.65

ters

1.63

TING

1.62

tered

1.62

ter

1.47

Activations Density 0.012%

words related to planned or intentional actions

past tense verbs, particularly those denoting actions or events

No Comments

No Known Activations