INDEX

Explanations

names and terms related to geopolitical conflicts and individuals involved in international issues

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

keye

-0.85

dayName

-0.77

taboola

-0.76

matically

-0.74

direction

-0.74

drivers

-0.74

words

-0.74

glers

-0.74

lance

-0.73

acles

-0.73

POSITIVE LOGITS

 Perez

1.16

 Berger

1.14

 Gonzalez

1.11

 Rodriguez

1.10

Luk

1.09

 Gonz

1.09

 Reed

1.09

 Martinez

1.09

 Berg

1.08

Sch

1.07

Activations Density 0.842%

names and terms related to geopolitical conflicts and individuals involved in international issues

No Comments

No Known Activations