INDEX

Explanations

words indicating uncertainty or probability

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res_post_32k-oai

Configuration

jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

32,768

Data Type

torch.float32

Hook Name

blocks.10.hook_resid_post

Hook Layer

Architecture

standard

Context Size

Dataset

Skylion007/openwebtext

Activation Function

topk

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 adolesc

-0.72

 srfAttach

-0.67

 Starts

-0.67

WithNo

-0.66

OTOS

-0.66

 Binding

-0.64

Reviewer

-0.62

CLASSIFIED

-0.62

 Flavoring

-0.62

Maker

-0.60

POSITIVE LOGITS

 deserve

1.35

owe

1.35

 have

1.30

 intend

1.29

 want

1.25

 belong

1.24

 reside

1.24

 prefer

1.17

 aren

1.17

 know

1.17

Activations Density 0.432%

words indicating uncertainty or probability

No Comments

No Known Activations