INDEX

Explanations

words related to initiating or beginning an action or discussion

oai_token-act-pair · gpt-3.5-turbo

the repeated word "let's" used 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

 Source

-0.72

è»

-0.70

eur

-0.69

 Reef

-0.68

CVE

-0.68

ELD

-0.67

 Able

-0.65

 restraints

-0.61

pak

-0.60

 Allied

-0.60

POSITIVE LOGITS

ween

0.91

acters

0.81

apo

0.76

emonic

0.76

wered

0.75

tha

0.75

 fulfill

0.75

 settle

0.74

nyder

0.74

achu

0.72

Activations Density 0.047%

words related to initiating or beginning an action or discussion

the repeated word "let's" used in various contexts

No Comments

No Known Activations