INDEX

Explanations

names of people

oai_token-act-pair · gpt-3.5-turbo

pronouns and names that indicate people or characters in a narrative

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

ourced

-0.70

 pregn

-0.64

ources

-0.64

 BASE

-0.63

 Dragonbound

-0.63

 fencing

-0.61

Redd

-0.60

VW

-0.60

 ILCS

-0.59

 runes

-0.59

POSITIVE LOGITS

auga

0.82

hower

0.72

arat

0.72

inite

0.68

amines

0.67

ghan

0.65

onen

0.64

nown

0.64

rag

0.63

teenth

0.62

Activations Density 0.129%

names of people

pronouns and names that indicate people or characters in a narrative

No Comments

No Known Activations