INDEX

Explanations

questions or inquiries directed towards understanding or obtaining information

oai_token-act-pair · gpt-3.5-turbo

inquiries and questions related to knowledge and information seeking

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

minus

-0.78

âĶĢâĶĢâĶĢâĶĢ

-0.72

----

-0.70

Statement

-0.70

apsed

-0.67

aganda

-0.66

ongs

-0.66

-0.65

 proud

-0.65

spoken

-0.64

POSITIVE LOGITS

 feasibility

0.82

 buyer

0.79

 legality

0.77

anian

0.76

 replacement

0.76

 revenge

0.74

 bids

0.74

 affordability

0.73

 possible

0.73

 revive

0.72

Activations Density 0.701%

questions or inquiries directed towards understanding or obtaining information

inquiries and questions related to knowledge and information seeking

No Comments

No Known Activations