INDEX

Explanations

quantitative terms related to amounts or numbers

oai_token-act-pair · gpt-3.5-turbo

references to amounts or quantities

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

iris

-0.70

SPONSORED

-0.69

nee

-0.65

Gate

-0.64

 braces

-0.64

pei

-0.60

Ã³n

-0.60

NJ

-0.59

opa

-0.58

 dads

-0.58

POSITIVE LOGITS

 quantity

4.04

 quantities

2.55

 Quantity

2.16

Quantity

1.94

quant

1.57

 volume

1.18

 amount

1.13

amount

1.13

 abundance

1.12

 potency

1.12

Activations Density 0.018%

quantitative terms related to amounts or numbers

references to amounts or quantities

No Comments

No Known Activations