INDEX

Explanations

questions being asked

oai_token-act-pair · gpt-3.5-turbo

references to questions being asked or answered

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ufact

-0.96

rites

-0.80

ensions

-0.73

ortunately

-0.69

axy

-0.69

orpor

-0.69

gradation

-0.68

 Tycoon

-0.68

alty

-0.67

ntil

-0.67

POSITIVE LOGITS

naires

1.37

naire

1.17

 answered

1.11

 questions

0.99

 posed

0.99

 pertaining

0.95

answered

0.90

 regarding

0.90

 asked

0.88

 unanswered

0.87

Activations Density 0.040%

questions being asked

references to questions being asked or answered

No Comments

No Known Activations