INDEX

Explanations

locations or places that have been visited by someone

oai_token-act-pair · gpt-3.5-turbo

instances of the word "visited" in various contexts

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

 stabilization

-0.65

ilt

-0.65

sure

-0.64

 Impro

-0.62

 fray

-0.61

 ignition

-0.58

Shield

-0.58

 Stall

-0.58

umph

-0.57

 actu

-0.57

POSITIVE LOGITS

 visited

3.60

 toured

1.97

 visits

1.94

 visit

1.84

 travelled

1.81

 traveled

1.77

 attended

1.76

 frequ

1.69

 inspected

1.68

 visiting

1.61

Activations Density 0.011%

locations or places that have been visited by someone

instances of the word "visited" in various contexts

No Comments

No Known Activations