INDEX

Explanations

adjectives related to evaluation or judgment

oai_token-act-pair · gpt-3.5-turbo

instances of products or events that are being evaluated for safety or viability

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

 Skydragon

-0.69

claw

-0.69

 FANTASY

-0.66

Zone

-0.62

Kit

-0.61

tackle

-0.60

 Pact

-0.60

Â¢

-0.59

joice

-0.58

CTV

-0.58

POSITIVE LOGITS

 itself

0.93

 oxid

0.74

 readable

0.70

 reboot

0.69

verty

0.65

 autonom

0.65

 smoothly

0.64

 dissolved

0.64

 veget

0.64

icably

0.64

Activations Density 0.703%

adjectives related to evaluation or judgment

instances of products or events that are being evaluated for safety or viability

No Comments

No Known Activations