INDEX

Explanations

words containing the subsequence "iar"

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals or characters

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

ester

-0.77

aic

-0.71

 Santos

-0.65

anwhile

-0.63

oak

-0.62

hend

-0.62

 Startup

-0.62

 Machina

-0.61

 Vend

-0.60

hetamine

-0.59

POSITIVE LOGITS

ror

1.06

osc

0.83

ibrary

0.79

æ©

0.78

bis

0.77

aries

0.75

arial

0.75

andom

0.74

ates

0.73

respective

0.72

Activations Density 0.028%

words containing the subsequence "iar"

references to specific individuals or characters

No Comments

No Known Activations