INDEX

Explanations

words related to a particular person or brand name, specifically "rod."

oai_token-act-pair · gpt-3.5-turbo

references to specific types of rods, likely in a scientific or biological context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 consultation

-0.66

 Tonight

-0.61

 Lonely

-0.60

 pleasing

-0.60

 partnership

-0.58

 entirety

-0.58

 Clare

-0.57

 consensus

-0.57

 totality

-0.56

 Pixie

-0.56

POSITIVE LOGITS

rod

1.61

ynamic

1.22

ynam

1.15

der

1.01

rig

1.00

lings

0.92

olph

0.89

bard

0.89

mund

0.85

roid

0.83

Activations Density 0.007%

words related to a particular person or brand name, specifically "rod."

references to specific types of rods, likely in a scientific or biological context

No Comments

No Known Activations