INDEX

Explanations

comparisons being made between different entities or concepts

oai_token-act-pair · gpt-3.5-turbo

instances of the word "comparison"

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

jong

-0.80

adia

-0.76

ignt

-0.75

der

-0.71

ieri

-0.70

msg

-0.69

ktop

-0.68

mic

-0.67

zyme

-0.67

 carpet

-0.66

POSITIVE LOGITS

isons

0.93

 comparisons

0.91

 favorably

0.90

 apples

0.81

 comparison

0.80

 comparing

0.79

 Compare

0.77

 compare

0.74

Osw

0.71

 Compar

0.69

Activations Density 0.017%

comparisons being made between different entities or concepts

instances of the word "comparison"

No Comments

No Known Activations