INDEX

Explanations

phrases related to examining, checking, or evaluating something

oai_token-act-pair · gpt-3.5-turbo

phrases indicating intention or actions aimed at verification and assessment

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Donation

-0.58

irection

-0.57

 fled

-0.57

eret

-0.56

living

-0.56

 belongs

-0.56

horn

-0.55

 fleeing

-0.54

 kneeling

-0.54

 Lives

-0.54

POSITIVE LOGITS

 determine

1.55

 ensure

1.35

 ascertain

1.33

see

1.30

 verify

1.22

 insure

1.20

 find

1.19

 discover

1.17

 gauge

1.09

 maximize

1.07

Activations Density 0.159%

phrases related to examining, checking, or evaluating something

phrases indicating intention or actions aimed at verification and assessment

No Comments

No Known Activations