INDEX

Explanations

instances of the word "fer" with various context and capitalization

oai_token-act-pair · gpt-3.5-turbo

references to official or government-related topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Decay

-0.72

ħĭ

-0.61

Bang

-0.60

 Improvement

-0.59

 havoc

-0.59

Products

-0.59

VP

-0.58

krit

-0.57

hift

-0.57

 invent

-0.57

POSITIVE LOGITS

rer

1.12

ior

1.05

ring

1.04

dinand

1.01

ocious

1.00

andom

0.99

rers

0.96

rets

0.91

ministic

0.91

rett

0.86

Activations Density 0.011%

instances of the word "fer" with various context and capitalization

references to official or government-related topics

No Comments

No Known Activations