INDEX

Explanations

keywords related to generalizations or inclusive statements

oai_token-act-pair · gpt-3.5-turbo

references to "everyone."

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

anga

-0.68

QL

-0.64

vern

-0.61

pose

-0.60

bal

-0.58

 testament

-0.57

 guiActiveUnfocused

-0.57

eeds

-0.56

eder

-0.56

duc

-0.56

POSITIVE LOGITS

 else

1.45

THING

1.03

 Else

0.99

 imaginable

0.93

else

0.92

Else

0.82

 conceivable

0.81

ãĤ§

0.77

who

0.76

igator

0.76

Activations Density 0.040%

keywords related to generalizations or inclusive statements

references to "everyone."

No Comments

No Known Activations