INDEX

Explanations

abbreviations and acronyms

oai_token-act-pair · gpt-3.5-turbo

verbs and prefixes that suggest action or transformation

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

SHIP

-0.86

Kah

-0.78

 Keller

-0.77

 Shots

-0.77

Tid

-0.75

 Halls

-0.71

lain

-0.70

 Bundy

-0.70

omsky

-0.69

sticks

-0.69

POSITIVE LOGITS

ighty

0.88

ertain

0.87

 plur

0.82

iring

0.81

ird

0.79

vent

0.78

 href

0.77

ploy

0.77

ception

0.77

cean

0.75

Activations Density 0.150%

abbreviations and acronyms

verbs and prefixes that suggest action or transformation

No Comments

No Known Activations