INDEX

Explanations

names of political figures from various parties and contexts

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

Compare

-0.74

yz

-0.65

 Lauder

-0.61

..................

-0.60

 Compare

-0.57

 Rothschild

-0.57

EVA

-0.56

PN

-0.55

LIN

-0.55

Pwr

-0.55

POSITIVE LOGITS

ELF

1.08

ullivan

1.05

own

1.04

 newest

0.92

selves

0.89

 favourite

0.87

 favorite

0.85

pecially

0.83

kaya

0.83

lightly

0.82

Activations Density 0.949%

names of political figures from various parties and contexts

No Comments

No Known Activations