INDEX

Explanations

references to allegations and accusations

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

keys

-0.87

arton

-0.77

 busiest

-0.76

tz

-0.74

å¸

-0.71

skill

-0.70

oppy

-0.69

IPS

-0.69

OTS

-0.69

patch

-0.68

POSITIVE LOGITS

 allegations

1.12

 alleging

1.09

 leveled

1.07

 levied

1.05

 accusing

1.05

 accusations

1.01

 accusation

1.01

 against

1.01

 allegation

0.99

 leve

0.98

Activations Density 0.041%

references to allegations and accusations

No Comments

No Known Activations