INDEX

Explanations

phrases describing positive outcomes or solutions

oai_token-act-pair · gpt-3.5-turbo

phrases indicating conditions or requirements that lead to something being considered favorable or better

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

soever

-0.85

phia

-0.76

 replaces

-0.66

 denies

-0.61

estic

-0.61

 Practices

-0.61

 Technologies

-0.60

 withdrew

-0.60

ALT

-0.60

 surrounds

-0.59

POSITIVE LOGITS

 interesting

1.18

 fascinating

1.09

 excellent

1.05

 awkward

1.05

interesting

1.05

 entertaining

1.04

 lousy

1.00

 tricky

0.98

 enjoyable

0.97

 intriguing

0.96

Activations Density 0.085%

phrases describing positive outcomes or solutions

phrases indicating conditions or requirements that lead to something being considered favorable or better

No Comments

No Known Activations