INDEX

Explanations

phrases related to people spending time or working together

oai_token-act-pair · gpt-3.5-turbo

instances of the word "together" and expressions of collaboration or teamwork

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ysis

-0.73

dark

-0.64

anty

-0.62

advant

-0.61

orers

-0.60

cock

-0.60

nit

-0.60

 Breach

-0.57

articles

-0.56

Fed

-0.56

POSITIVE LOGITS

ieth

0.69

halla

0.69

å¥³

0.68

!.

0.67

standing

0.65

 impunity

0.64

heid

0.64

inian

0.64

atives

0.64

ingen

0.63

Activations Density 0.023%

phrases related to people spending time or working together

instances of the word "together" and expressions of collaboration or teamwork

No Comments

No Known Activations