INDEX

Explanations

words related to compounds or composites

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "comp" or related variations, primarily indicating compounding concepts or behaviors

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Millennium

-0.77

 Sandwich

-0.71

 Rabbit

-0.70

 Roses

-0.68

 Spectrum

-0.68

Alm

-0.67

 Merry

-0.65

 Variety

-0.65

 Lumpur

-0.64

FORE

-0.64

POSITIVE LOGITS

endium

1.32

ulsive

1.29

artments

1.16

ilers

1.08

ressing

1.05

acts

1.05

rehens

1.04

ounding

1.03

otent

1.02

osit

0.99

Activations Density 0.011%

words related to compounds or composites

occurrences of the term "comp" or related variations, primarily indicating compounding concepts or behaviors

No Comments

No Known Activations