INDEX

Explanations

numerical expressions indicating an almost amount or close approximation

oai_token-act-pair · gpt-3.5-turbo

instances of the word "almost" followed by numerical values or quantities

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

oris

-0.76

agate

-0.75

Reviewer

-0.71

ieu

-0.70

illary

-0.70

 bearer

-0.69

alam

-0.67

DH

-0.67

venant

-0.67

arten

-0.66

POSITIVE LOGITS

 identical

0.73

capacity

0.69

 ident

0.67

0.64

 unprecedented

0.64

 saturated

0.64

zero

0.64

 electr

0.64

 unrelated

0.64

Activations Density 0.039%

numerical expressions indicating an almost amount or close approximation

instances of the word "almost" followed by numerical values or quantities

No Comments

No Known Activations