INDEX

Explanations

ratios or comparisons between different quantities or values

oai_token-act-pair · gpt-3.5-turbo

phrases that mention ratios 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

authorized

-0.68

 criminal

-0.67

CAN

-0.63

Explore

-0.61

help

-0.61

criminal

-0.61

 Holden

-0.59

 ENTER

-0.59

EGIN

-0.58

DEC

-0.58

POSITIVE LOGITS

 ratio

4.08

 ratios

3.46

 Ratio

3.26

 proportion

1.64

 proportions

1.59

 percentage

1.36

 rate

1.36

 density

1.26

 ration

1.24

 imbalance

1.19

Activations Density 0.008%

ratios or comparisons between different quantities or values

phrases that mention ratios in various contexts

No Comments

No Known Activations