INDEX

Explanations

keywords related to total values or quantities

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "total" and related numerical summaries

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

anx

-0.73

 mold

-0.68

 welcoming

-0.67

 manner

-0.66

lder

-0.64

 teaching

-0.63

 fashion

-0.63

 emer

-0.62

 anch

-0.62

 inform

-0.62

POSITIVE LOGITS

Total

3.90

total

2.54

 Total

2.28

 TOTAL

1.99

 total

1.63

OTAL

1.61

Average

1.58

Percent

1.47

Overall

1.41

Number

1.37

Activations Density 0.016%

keywords related to total values or quantities

occurrences of the word "total" and related numerical summaries

No Comments

No Known Activations