INDEX

Explanations

phrases related to topics or concepts

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate association or connection to a subject

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

ashtra

-0.72

venants

-0.69

dit

-0.69

etooth

-0.67

daq

-0.65

tackle

-0.62

eca

-0.62

itored

-0.62

rys

-0.61

renheit

-0.60

POSITIVE LOGITS

 consists

0.64

 comprises

0.62

pires

0.61

 lacks

0.57

reads

0.57

 Collider

0.56

oln

0.55

 depends

0.55

 softened

0.54

 change

0.54

Activations Density 0.557%

phrases related to topics or concepts

phrases that indicate association or connection to a subject

No Comments

No Known Activations