INDEX

Explanations

occurrences of repeated phrases or terms throughout the document

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

 unbeliev

-0.75

 AMERICA

-0.74

ãĤ´ãĥ³

-0.73

 exha

-0.68

 adolesc

-0.68

FTWARE

-0.68

 referen

-0.67

âĸ¬

-0.66

hips

-0.66

é¾

-0.65

POSITIVE LOGITS

ublic

1.20

rint

1.17

isodes

1.13

alm

1.10

lex

1.07

osit

1.06

stakes

1.05

edia

1.05

artment

1.01

oleon

1.00

Activations Density 0.009%

occurrences of repeated phrases or terms throughout the document

No Comments

No Known Activations