INDEX

Explanations

official titles or positions held by individuals

oai_token-act-pair · gpt-3.5-turbo

terms related to important choices or options

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

hen

-0.68

cius

-0.68

sv

-0.62

visible

-0.61

athi

-0.61

 Gleaming

-0.61

zek

-0.60

Dash

-0.59

ku

-0.58

POSITIVE LOGITS

for

1.22

for

1.09

FOR

1.01

FOR

0.96

For

0.89

For

0.85

abase

0.69

fore

0.68

fort

0.67

 against

0.67

Activations Density 0.175%

official titles or positions held by individuals

terms related to important choices or options

No Comments

No Known Activations