INDEX

Explanations

adjectives related to quantity or degree

oai_token-act-pair · gpt-3.5-turbo

comparative terms indicating high and low 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 @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

sov

-0.79

dor

-0.75

irit

-0.69

)</

-0.63

orio

-0.63

ools

-0.61

RIPT

-0.56

RAW

-0.56

ador

-0.56

rique

-0.55

POSITIVE LOGITS

 enough

1.13

 compared

1.03

(>

1.03

(<

0.94

 relative

0.90

(~

0.87

enough

0.82

 proportion

0.81

 earners

0.78

est

0.78

Activations Density 0.097%

adjectives related to quantity or degree

comparative terms indicating high and low values or quantities

No Comments

No Known Activations