INDEX

Explanations

phrases related to events or actions happening during a specific time frame

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

coins

-0.88

ettle

-0.77

cko

-0.73

 guessed

-0.72

ographically

-0.67

onica

-0.67

usercontent

-0.66

 prevailed

-0.66

cloth

-0.65

 Cosponsors

-0.65

POSITIVE LOGITS

 renovations

1.20

 festivities

1.09

 deliberations

1.07

 negotiations

1.07

 trials

1.05

 filming

1.05

 session

1.04

 renovation

1.04

 sessions

1.03

 trial

1.00

Activations Density 5.498%

phrases related to events or actions happening during a specific time frame

No Comments

No Known Activations