INDEX

Explanations

phrases or names containing the word "Hem"

oai_token-act-pair · gpt-3.5-turbo

references to characters or individuals with the name "Hem."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 deaf

-0.67

 jewelry

-0.62

Brand

-0.62

 sterling

-0.61

 concerted

-0.60

 recess

-0.60

 Richie

-0.59

 unauthorized

-0.59

ãĥł

-0.58

ãģį

-0.57

POSITIVE LOGITS

Hem

1.16

sworth

0.94

oglobin

0.93

odynamics

0.87

oran

0.87

obiles

0.86

icide

0.83

neum

0.83

grim

0.83

borgh

0.82

Activations Density 0.013%

phrases or names containing the word "Hem"

references to characters or individuals with the name "Hem."

No Comments

No Known Activations