INDEX

Explanations

words related to being absorbed or immersed in something

oai_token-act-pair · gpt-3.5-turbo

references to being confined or restricted, often metaphorically represented by the word "bed."

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

IGH

-0.66

 strateg

-0.60

 warr

-0.60

å¦

-0.59

pac

-0.57

 foresee

-0.57

 Phill

-0.56

umer

-0.56

Ul

-0.55

 import

-0.55

POSITIVE LOGITS

bed

4.79

bing

2.58

bed

1.59

 beds

1.55

Bed

1.47

bered

1.40

bers

1.36

Bed

1.36

ber

1.33

bs

1.30

Activations Density 0.005%

words related to being absorbed or immersed in something

references to being confined or restricted, often metaphorically represented by the word "bed."

No Comments

No Known Activations