INDEX

Explanations

medical conditions or emergencies

oai_token-act-pair · gpt-3.5-turbo

terms associated with acute conditions or experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

wagen

-0.97

essee

-0.80

hov

-0.79

pherd

-0.78

vous

-0.77

hops

-0.73

yards

-0.73

ships

-0.73

tz

-0.72

ertodd

-0.71

POSITIVE LOGITS

 coronary

0.85

 acute

0.84

 exacerb

0.79

 respiratory

0.78

 poisoning

0.76

 distress

0.73

 sensitivity

0.73

ptoms

0.72

entimes

0.71

 shortage

0.71

Activations Density 0.007%

medical conditions or emergencies

terms associated with acute conditions or experiences

No Comments

No Known Activations