INDEX

Explanations

adjectives indicating quantity or degree such as "much," "a lot," and "so much."

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ares

-0.85

aves

-0.84

breakers

-0.79

Ts

-0.75

buster

-0.75

ends

-0.73

runs

-0.73

save

-0.73

agents

-0.72

iques

-0.71

POSITIVE LOGITS

 overlap

1.17

 evidence

1.05

 disagreement

1.04

 confusion

1.03

 similarity

1.03

 uncertainty

1.00

 discrepancy

0.99

 ambiguity

0.98

 possibility

0.98

 indication

0.97

Activations Density 0.136%

adjectives indicating quantity or degree such as "much," "a lot," and "so much."

No Comments

No Known Activations