INDEX

Explanations

Twitter handles preceded by "@" symbol

oai_token-act-pair · gpt-3.5-turbo

social media handles or mentions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 specificity

-0.90

 Sloven

-0.76

 exact

-0.76

 grasp

-0.76

 influenza

-0.74

 encaps

-0.73

 vagina

-0.73

 maternity

-0.73

 crystall

-0.72

 topping

-0.72

POSITIVE LOGITS

john

1.33

thereal

1.30

maxwell

1.29

bryce

1.21

jac

1.21

mr

1.19

Coach

1.18

mc

1.17

jon

1.16

Real

1.14

Activations Density 0.032%

Twitter handles preceded by "@" symbol

social media handles or mentions

No Comments

No Known Activations