INDEX

Explanations

phrases indicating something is being sacrificed or negatively impacted in favor of something else

oai_token-act-pair · gpt-3.5-turbo

references to sacrifices or costs incurred for the benefit of others

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

odes

-0.75

 Aval

-0.68

arre

-0.67

DIT

-0.67

kens

-0.64

 Entry

-0.61

 fuzz

-0.61

ynt

-0.60

Nig

-0.60

gran

-0.60

POSITIVE LOGITS

 expense

1.05

 spared

0.84

iasis

0.81

 incurred

0.80

liest

0.80

 altar

0.78

ãĥĨãĤ£

0.74

llo

0.72

 taxpayers

0.72

rophe

0.71

Activations Density 0.011%

phrases indicating something is being sacrificed or negatively impacted in favor of something else

references to sacrifices or costs incurred for the benefit of others

No Comments

No Known Activations