INDEX

Explanations

phrases related to step-by-step instructions or process descriptions

oai_token-act-pair · gpt-3.5-turbo

second-person references in various contexts

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

ðĿ

-0.71

à

-0.68

OX

-0.64

idity

-0.63

abases

-0.62

 Unreal

-0.59

istries

-0.58

 Valiant

-0.58

CHAT

-0.58

fff

-0.57

POSITIVE LOGITS

ensis

0.74

 progresses

0.74

 progressed

0.72

 relates

0.72

 transitioned

0.71

 matured

0.71

 unfolds

0.70

 regards

0.67

 evolves

0.64

soDeliveryDate

0.64

Activations Density 0.190%

phrases related to step-by-step instructions or process descriptions

second-person references in various contexts

No Comments

No Known Activations