INDEX

Explanations

the word "which" followed by a number

oai_token-act-pair · gpt-3.5-turbo

the word "which" in various contexts indicating a focus on specific clauses or examples

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

rolet

-0.73

ctor

-0.71

Problem

-0.70

ifest

-0.67

strap

-0.64

et

-0.63

Typ

-0.62

ÂłÂłÂłÂłÂłÂłÂłÂł

-0.62

unch

-0.61

ct

-0.60

POSITIVE LOGITS

soever

0.99

 guts

0.74

xual

0.73

ĸļ

0.71

upon

0.70

adoes

0.68

[|

0.68

 case

0.66

ornia

0.65

andom

0.65

Activations Density 0.041%

the word "which" followed by a number

the word "which" in various contexts indicating a focus on specific clauses or examples

No Comments

No Known Activations