INDEX

Explanations

words related to trademarks or products

oai_token-act-pair · gpt-3.5-turbo

terms related to academic positions or classifications

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

IRD

-0.74

 DRAG

-0.70

 srfAttach

-0.66

galitarian

-0.66

isSpecialOrderable

-0.66

 Tunis

-0.65

OE

-0.64

 mosqu

-0.62

abwe

-0.62

ESH

-0.61

POSITIVE LOGITS

ures

1.04

ure

0.93

heid

0.87

ions

0.85

icularly

0.83

umbn

0.82

0.77

rontal

0.76

baugh

0.74

cil

0.74

Activations Density 0.087%

words related to trademarks or products

terms related to academic positions or classifications

No Comments

No Known Activations