INDEX

Explanations

instances where something is being proven or demonstrated

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "proving"

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

scribe

-0.71

intend

-0.68

 evacuate

-0.67

irst

-0.67

uci

-0.64

abin

-0.62

isson

-0.62

chairs

-0.59

Ã±

-0.57

escent

-0.57

POSITIVE LOGITS

 proving

3.54

 demonstrating

1.92

 showing

1.57

 verifying

1.54

 proof

1.48

 confirming

1.47

 illustrating

1.43

 showcasing

1.41

 documenting

1.34

 admitting

1.33

Activations Density 0.011%

instances where something is being proven or demonstrated

occurrences of the word "proving"

No Comments

No Known Activations