INDEX

Explanations

phrases indicating something is yet to be confirmed or announced

oai_token-act-pair · gpt-3.5-turbo

phrases indicating pending or unresolved situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Arrows

-0.71

 Lives

-0.69

 Failure

-0.68

 srfAttach

-0.67

 Lose

-0.67

 Falling

-0.67

 Waste

-0.65

 Worst

-0.65

 Akin

-0.64

 Fail

-0.64

POSITIVE LOGITS

 ascert

1.44

 announced

1.37

 determined

1.35

 finalized

1.30

 revealed

1.28

 decided

1.25

 disclosed

1.25

 confirmed

1.21

 unveiled

1.13

 clarified

1.12

Activations Density 0.128%

phrases indicating something is yet to be confirmed or announced

phrases indicating pending or unresolved situations

No Comments

No Known Activations