INDEX

Explanations

phrases related to prominent people or events

oai_token-act-pair · gpt-3.5-turbo

proper nouns and specific terms related to entities or events

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 princ

-0.78

 decl

-0.72

 subp

-0.71

 Princ

-0.70

 McGu

-0.69

mos

-0.68

 Chap

-0.68

apr

-0.68

Uni

-0.67

 hemor

-0.67

POSITIVE LOGITS

ing

1.57

ings

1.28

ed

1.27

ING

1.09

edIn

1.07

edly

1.05

ington

1.01

ment

1.00

ization

0.98

ership

0.98

Activations Density 0.165%

phrases related to prominent people or events

proper nouns and specific terms related to entities or events

No Comments

No Known Activations