INDEX

Explanations

detailed analyses or reviews

oai_token-act-pair · gpt-3.5-turbo

phrases related to analysis or evaluations of various subjects

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

ulla

-0.78

KK

-0.67

ription

-0.67

ittees

-0.66

Pg

-0.65

ertodd

-0.65

Lib

-0.65

dayName

-0.63

ktop

-0.63

 teasp

-0.62

POSITIVE LOGITS

 course

0.87

 sorts

0.86

 Excellence

0.71

issance

0.71

tnc

0.67

how

0.66

 behalf

0.64

articles

0.63

ussian

0.61

 rationality

0.61

Activations Density 0.158%

detailed analyses or reviews

phrases related to analysis or evaluations of various subjects

No Comments

No Known Activations