INDEX

Explanations

phrases related to public services and policies

oai_token-act-pair · gpt-3.5-turbo

mentions of public service and related organizations or announcements

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

heric

-0.72

ularity

-0.71

agher

-0.70

Pacific

-0.70

jad

-0.68

nen

-0.67

ewski

-0.65

ppy

-0.64

wered

-0.62

Lago

-0.61

POSITIVE LOGITS

!--

0.73

ilts

0.66

 servant

0.63

+++

0.63

oppable

0.62

pection

0.61

ufact

0.60

 announcements

0.60

 toll

0.59

 servants

0.59

Activations Density 0.092%

phrases related to public services and policies

mentions of public service and related organizations or announcements

No Comments

No Known Activations