INDEX

Explanations

phrases related to government organizations, especially agencies dealing with food and drugs

oai_token-act-pair · gpt-3.5-turbo

mentions of food and drug regulatory agencies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

pole

-0.77

 unfocusedRange

-0.70

rontal

-0.69

 Whale

-0.65

IDENT

-0.64

Dutch

-0.62

onde

-0.62

ciating

-0.61

 Penguin

-0.61

Ic

-0.59

POSITIVE LOGITS

 beverage

1.05

 beverages

1.05

 drink

1.02

 Drink

0.91

 nutrition

0.90

 nutritious

0.86

thood

0.80

illin

0.79

 drinks

0.79

 thirsty

0.78

Activations Density 0.194%

phrases related to government organizations, especially agencies dealing with food and drugs

mentions of food and drug regulatory agencies

No Comments

No Known Activations