INDEX

Explanations

mentions of locations and events, particularly in the context of legal proceedings

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

acters

-0.66

orically

-0.63

RL

-0.60

OULD

-0.59

Â¯

-0.58

ults

-0.57

ELF

-0.56

UTF

-0.56

ython

-0.55

ãĤ§

-0.55

POSITIVE LOGITS

stage

1.08

 board

1.04

ibaba

1.02

shore

1.00

 behalf

0.99

erous

0.99

etime

0.95

site

0.94

 rooft

0.92

demand

0.92

Activations Density 0.470%

mentions of locations and events, particularly in the context of legal proceedings

No Comments

No Known Activations