INDEX

Explanations

words related to overflow or excess

oai_token-act-pair · gpt-3.5-turbo

terms related to overflow or excess effects

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

#$

-0.72

YES

-0.65

 Thur

-0.64

killed

-0.63

 Samp

-0.60

 Ridley

-0.60

 Desmond

-0.59

crow

-0.59

Jed

-0.58

 Fritz

-0.58

POSITIVE LOGITS

onomous

0.82

omnia

0.75

anced

0.73

DragonMagazine

0.71

antly

0.68

jriwal

0.68

tnc

0.68

orest

0.67

iste

0.67

itionally

0.66

Activations Density 0.296%

words related to overflow or excess

terms related to overflow or excess effects

No Comments

No Known Activations