INDEX

Explanations

words related to using tools or resources to accomplish a task efficiently

oai_token-act-pair · gpt-3.5-turbo

words and concepts related to data usage and security issues

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

ategories

-0.72

bernatorial

-0.70

ilogy

-0.70

 subsequ

-0.70

ensable

-0.68

ufact

-0.67

ategory

-0.67

pleted

-0.67

irement

-0.66

icipated

-0.65

POSITIVE LOGITS

 sparing

0.99

 wisely

0.86

 levers

0.80

 pseudonym

0.75

 analogy

0.74

 gimm

0.74

 liber

0.73

 tools

0.71

 metaphors

0.70

 metaphor

0.68

Activations Density 0.483%

words related to using tools or resources to accomplish a task efficiently

words and concepts related to data usage and security issues

No Comments

No Known Activations