INDEX

Explanations

phrases related to events and activities happening in a sequence

oai_token-act-pair · gpt-3.5-turbo

references to repeated events or conditions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 diarr

-0.77

 emphas

-0.74

 realise

-0.73

 scept

-0.70

 behaviour

-0.68

 practise

-0.66

 rubbish

-0.65

 doesnt

-0.65

 realised

-0.65

 labour

-0.64

POSITIVE LOGITS

ãĤ´ãĥ³

0.84

Saturday

0.66

aukee

0.65

onductor

0.64

tained

0.63

 onstage

0.62

luster

0.62

Lex

0.62

heastern

0.62

EMBER

0.61

Activations Density 0.706%

phrases related to events and activities happening in a sequence

references to repeated events or conditions

No Comments

No Known Activations