INDEX

Explanations

words related to specific names, particularly "Helen" with a high activation level compared to other names

oai_token-act-pair · gpt-3.5-turbo

occurrences of the name "Helen" and related prominent individuals

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

inal

-0.88

arette

-0.88

awaru

-0.87

emonium

-0.82

endar

-0.81

arettes

-0.81

ileaks

-0.80

entin

-0.80

agne

-0.77

otine

-0.76

POSITIVE LOGITS

ship

0.85

sea

0.74

locked

0.70

nai

0.69

 Sloan

0.67

 Keller

0.67

housing

0.67

sworth

0.67

locks

0.67

tox

0.67

Activations Density 0.042%

words related to specific names, particularly "Helen" with a high activation level compared to other names

occurrences of the name "Helen" and related prominent individuals

No Comments

No Known Activations