INDEX

Explanations

phrases related to legal or medical terms

oai_token-act-pair · gpt-3.5-turbo

mentions of specific medical or educational evaluations and involuntary actions

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

sonian

-0.82

 litter

-0.75

icians

-0.68

-0.66

Kings

-0.65

 veto

-0.64

ically

-0.64

 adjourn

-0.63

ãĥ£

-0.62

geo

-0.61

POSITIVE LOGITS

EE

1.09

ã

0.81

terness

0.79

advant

0.79

lections

0.78

ossession

0.77

luent

0.75

irmed

0.74

Pwr

0.74

awei

0.74

Activations Density 0.024%

phrases related to legal or medical terms

mentions of specific medical or educational evaluations and involuntary actions

No Comments

No Known Activations