INDEX

Explanations

phrases related to user interaction and interfaces

oai_token-act-pair · gpt-3.5-turbo

terms related to user experience and customer interaction

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ĪĴ

-0.84

amazon

-0.75

ropolitan

-0.74

 veins

-0.74

Reloaded

-0.67

GMT

-0.66

Wol

-0.64

leck

-0.64

Ö

-0.64

rontal

-0.63

POSITIVE LOGITS

ause

0.68

groups

0.67

poses

0.66

jac

0.66

 Cosponsors

0.65

escription

0.64

atures

0.63

âĦ¢:

0.63

generated

0.61

amental

0.61

Activations Density 0.367%

phrases related to user interaction and interfaces

terms related to user experience and customer interaction

No Comments

No Known Activations