INDEX

Explanations

terms related to PG ratings

oai_token-act-pair · gpt-3.5-turbo

mentions of the term "PG" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Horizons

-0.74

Ic

-0.71

joined

-0.68

advert

-0.67

hold

-0.67

self

-0.66

ker

-0.64

adm

-0.61

section

-0.60

introdu

-0.60

POSITIVE LOGITS

PG

1.34

eln

0.89

Cs

0.87

VG

0.86

DF

0.86

osate

0.85

PI

0.85

ENE

0.84

RAM

0.83

PF

0.82

Activations Density 0.007%

terms related to PG ratings

mentions of the term "PG" in various contexts

No Comments

No Known Activations