INDEX

Explanations

instances of the word "go" in various contexts

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

lake

-0.67

picking

-0.65

 ammon

-0.64

ament

-0.64

mith

-0.63

race

-0.62

 Aram

-0.61

urgy

-0.61

role

-0.59

lain

-0.59

POSITIVE LOGITS

vernment

1.07

verning

1.04

ven

0.96

ffic

0.92

ogly

0.87

lems

0.84

etz

0.83

zzi

0.83

zzo

0.83

ppe

0.80

Activations Density 0.008%

instances of the word "go" in various contexts

No Comments

No Known Activations