INDEX

Explanations

words related to legal or official matters

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate a connection or relevance to 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

ifter

-0.81

âķĲ

-0.78

aneers

-0.73

 Muller

-0.73

imates

-0.69

UGE

-0.69

ardless

-0.68

arer

-0.67

 Trees

-0.67

âķĲâķĲ

-0.66

POSITIVE LOGITS

 thereto

0.92

 ancest

0.88

worldly

0.86

paren

0.79

 unrelated

0.77

lly

0.77

ness

0.77

 reconc

0.74

rums

0.73

ciplinary

0.73

Activations Density 0.018%

words related to legal or official matters

phrases that indicate a connection or relevance to various subjects

No Comments

No Known Activations