INDEX

Explanations

phrases related to a close examination or scrutiny

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "look" 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 expulsion

-0.61

 enrichment

-0.60

 together

-0.60

ffen

-0.60

 solar

-0.59

em

-0.59

 epis

-0.57

 Concord

-0.56

 colonization

-0.56

 serving

-0.55

POSITIVE LOGITS

look

4.47

looking

1.74

Look

1.61

 LOOK

1.55

 Look

1.42

 look

1.36

come

1.31

 looks

1.22

Looks

1.13

stand

1.13

Activations Density 0.008%

phrases related to a close examination or scrutiny

occurrences of the word "look" in various contexts

No Comments

No Known Activations