INDEX

Explanations

words related to treating or treatment

oai_token-act-pair · gpt-3.5-turbo

terms related to treatment

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 discrepancies

-0.67

ãģ®éŃĶ

-0.63

å£«

-0.62

MSM

-0.61

 pending

-0.60

 offline

-0.59

 Clay

-0.59

 verified

-0.59

 overtime

-0.58

 garbage

-0.58

POSITIVE LOGITS

reat

4.45

reath

2.10

reating

1.75

 Treat

1.27

reated

1.11

ree

1.07

 Breat

1.02

 breat

0.95

utic

0.94

irds

0.92

Activations Density 0.007%

words related to treating or treatment

terms related to treatment

No Comments

No Known Activations