INDEX

Explanations

words containing the pattern "fe" followed by a number, which is likely related to some specific context or term

oai_token-act-pair · gpt-3.5-turbo

the presence of the word "fe" in various contexts

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

 Nadu

-0.73

uyomi

-0.71

 Redux

-0.69

 Shining

-0.68

İĭ

-0.67

 Crunch

-0.67

 Dominion

-0.67

nian

-0.66

ova

-0.66

DOWN

-0.64

POSITIVE LOGITS

ldom

1.06

vered

1.03

cking

1.02

ld

1.01

eling

1.00

gging

0.97

fe

0.94

ences

0.93

eking

0.92

ete

0.92

Activations Density 0.010%

words containing the pattern "fe" followed by a number, which is likely related to some specific context or term

the presence of the word "fe" in various contexts

No Comments

No Known Activations