INDEX

Explanations

phrases related to negation or denial of some kind

oai_token-act-pair · gpt-3.5-turbo

terms related to medical necessity and documentation

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

ortment

-0.72

tnc

-0.69

haul

-0.67

éĹĺ

-0.66

incial

-0.63

rahim

-0.62

 ACTIONS

-0.62

href

-0.61

YES

-0.61

SAN

-0.60

POSITIVE LOGITS

 anymore

1.77

nor

1.49

 necessarily

1.08

 anywhere

1.00

 whatsoever

0.97

 slightest

0.97

yet

0.92

yet

0.88

 anybody

0.86

 anything

0.86

Activations Density 0.410%

phrases related to negation or denial of some kind

terms related to medical necessity and documentation

No Comments

No Known Activations