INDEX

Explanations

phrases related to exploring or investigating a topic or subject in depth

oai_token-act-pair · gpt-3.5-turbo

terms related to diving or delving into subjects or topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

minus

-0.68

ouf

-0.64

 redistributed

-0.64

 Proud

-0.61

quad

-0.60

 projection

-0.60

uv

-0.60

iop

-0.59

weather

-0.59

 rally

-0.59

POSITIVE LOGITS

 deeper

1.39

 into

1.26

 Into

1.04

 INTO

1.02

 deeply

0.98

 deep

0.98

 deepest

0.95

 dive

0.90

into

0.85

 delve

0.85

Activations Density 0.090%

phrases related to exploring or investigating a topic or subject in depth

terms related to diving or delving into subjects or topics

No Comments

No Known Activations