INDEX

Explanations

comparisons or evaluations followed by a positive or negative sentiment

oai_token-act-pair · gpt-3.5-turbo

phrases that assert a state or identity

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

itches

-0.70

ffe

-0.70

actory

-0.65

atre

-0.64

ees

-0.62

ievers

-0.61

gg

-0.61

lex

-0.60

Gir

-0.60

agon

-0.60

POSITIVE LOGITS

 admittedly

1.02

 comprised

0.94

 basically

0.92

 supposed

0.91

 essentially

0.90

 meant

0.89

 unlikely

0.88

 undoubtedly

0.87

 presumably

0.86

 obviously

0.86

Activations Density 0.164%

comparisons or evaluations followed by a positive or negative sentiment

phrases that assert a state or identity

No Comments

No Known Activations