INDEX

Explanations

references to the FOX network and its content

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

acca

-0.84

ucket

-0.81

lain

-0.76

 Edison

-0.76

ainer

-0.74

adding

-0.71

aca

-0.69

ivari

-0.67

accompan

-0.67

abee

-0.67

POSITIVE LOGITS

FOX

1.10

 SPORTS

1.05

CAST

1.01

TAG

0.92

DEF

0.91

conn

0.91

WOOD

0.88

PRO

0.88

DEN

0.85

IRO

0.84

Activations Density 0.005%

references to the FOX network and its content

No Comments

No Known Activations